Ahmadzei's picture
added 3 more tables for large emb model
5fa1a76
This does not
apply to models leveraging the Encoder-Decoder framework.